E. A. TH.VERDURMEN
678
Mechanism of Oxygen Isotopic Exchange in Mixtures of Carbon Monoxide and Oxygen in a Quartz Vessel
by E. A. Th. Verdurmen F.O.M.--lnstitiLut
toor Atoom, en Molecuulfysica, Amsterdam, T h e iyetherlands
(Receiz’ed October 3, 1966)
The mechanism of the oxygen isotopic exchange processes occurring in a thermally reacting dry mixture of natural CO and 02,predominantly enriched in the l 8 0 1 * 0 species, has been studied between 500 and 580” and within the pressure range 28-144 torr, in a quartz reaction vessel. Three independent processes occur: (1) formation of CO,; (2) oxygen isotopic exchange between CO and 0,; and (3) scrambling of oxygen isotopes in 0,. Scrambling and (30-02 isotopic exchange are retarded practically equally by the addition of considerable amounts of foreign gases, He, Ar, Xe, Sz,COZ, and SFo, whereas the rate of CO, formation remains practically constant. Relative efficiencies of the various foreign gases for retardation of the exchange processes have been obtained. Both scrambling and CO-0, isotopic exchange are completely inhibited by the addition of small amounts of BO. The processes involved show similar apparent activation energies between 30 and 35 kcal/mole. The formation of CO, is mainly a surface process. Scrambling and CO-02 isotopic exchange are interpreted in terms of gas-phase chain reactions. The chain carriers are oxygen atoms, originating from a wall reaction between CO and 02, and excited CO, (singlet) molecules formed by radiationless transition from the COZ (triplet) association complex of CO and 0. Chain termination occurs both by wall recombination of oxygen atoms and by gas-phase third-order reactions. The CO-02 isotopic exchange is determined by the reactions 0 CO + CO2* (triplet) + CO 0, and scrambling is ascribed to 0 O2 + 03*+ 0, 0 and COz** (singlet) 0 2 C02* (triplet) 0,. S o marked deactivation of excited C o n has been observed. Addition of foreign gases favors third-order gas-phase termination and consequently reduces the 0 atom partial pressure. The inhibitory effect of XO is ascribed to very small amounts of NO,, formed by slow reaction of K O with the O2 reactant, and acting as an 0 atom scavenger. The ratio of the rate constant kz5of 0 02 + Os* and the rate constant k28 of 0 CO + COS” (triplet) has been obtained at 500”, k25/k28 = 7.1 0.2.
+
+
+
+
Introduction Generally, the study of isotopic exchange reactions can demonstrate the existence of otherwise nondetectable processes, or can provide information on the elementary steps of a known but complicated chemical process. Such a complicated system is a thermally reacting dry mixture of CO and 0 2 . It exhibits the phenomenon of the first and second explosion limits.lJ The explosion is of the chain-branching type. Oxygen atoms and excited COz (triplet) molecules3 next to the carbon suboxides,* CzO and C302, are proposed as the chain carriers or as important intermediates. The T h e Journal of Phgsical Chemistry
+
*
-
+
+
+
explosion peninsula starts above 600°.5 Below that temperature, a slow formation of CO, is observed. In a previous paper,6 it has been reported that in addition, (1) G. J. Minkoff and C. F. H. Tipper, “Chemistry of Combustion Reactions,” Butterworth and Co. (Publishers) Ltd., London, 1962, p 58. (2) B. Lewis and G. von Elbe, “Combustion, Flames and Explosions of Gases,” Academic Press Inc., New York, N. Y., 1961, p 71. (3) S. W. Benson, “The Foundations of Chemical Kinetics,” McGraw-Hill Book Co., Inc., New York, N. Y., 1960, p 460. (4) P. Harteck and S.Dondes, J . Chem. Phys., 27, 1419 (1957). (5) P. G. Dickens, J. E. Dove, and J. W. Linnett, Trans. Faraday SOC.,60, 539 (1964).
MECHANISM OF OXYGEN ISOTOPIC EXCHANGE IN CO-02 MIXTURES
above 450", oxygen isotopic exchange occurs between CO and O2 as an independent gas-phase process, whereas the attendant formation of C02 is mainly a surface process. The preliminary study involved left the following problems unsolved : (a) are the intermediates mentioned, and particularly are oxygen atoms important in the exchange mechanism; (b) is a possibly formed excited C02 molecule sufficiently long lived to take part in the reactions; (c) does oxygen scrambling in O2 occur (02-O2 isotopic exchange); (d) if oxygen atoms are important intermediates, what is their origin? The present paper reports the results of an attempt to obtain a more complete understanding of the processes occurring in a slowly and thermally reacting dry7 mixture of CO and 0 2 in a quartz vessel. The oxygen isotopic exchange between the reactants and the formation of CO, have been followed directly and simultaneously as a function of reaction time by continuous analysis of reactants and products with a mass spectrometer. The results indicate that oxygen atoms act as chain carriers in the oxygen exchange between the reactants.
Experimental Section Apparatus. The experiments were performed in a reaction vessel constructed of Vitreosil quartz, i.d. 39 mm, volume 200 cc, supplied with a small quartz leak (3 X 10-5 1. torr sec-l, a t a pressure difference of 75 torr and at room temperature) to the ionization chamber of an Atlas CH4 mass spectrometer. As shown in Figure 1, the reaction vessel was placed in a resistance-heated furnace consisting of a copper cylinder, tightly surrounding the vessel and provided with external windings, and heat-insulating material. The furnace temperature was maintained constant within ~k0.5"at 500" by means of a temperature controller. Control was applied to the furnace current by the changing resistance of a P t resistance thermometer, fitted into the copper cylinder as indicated in Figure 1. The temperature gradient along the reaction vessel was less than 2" at 500". Temperatures were measured by thermocouples (chromel-alumel) fitted outside the reaction vessel and inside the copper cylinder. The reactor was separated from the vacuum line by an indium stopcock. This particular vacuum line and the attendant pumping system were used exclusively for the evacuation of the reactor. The reactants inlet at the reactor side consisted of about 100 cm of capillary tube, i.d. 2 mm, partly cooled at -78" to prevent interference of condensable impurities (e.g., mercury vapor from the manometer and the Toepler pump). The inlet system consisted of a 500-cc vessel for pre-
679
E
e
\
"
- G
-1
1-
33 0 mm
---------- --
x--; p:-5:: 10" torr
pmm
10 mm
~
LIS
Figure 1. The reactor and its connections; the quartz leak. A, external windings; B, heat-insulating material; C, copper cylinder; D, Vitreosil; E, indium stopcock; F, to vacuum line I; GI reactants inlet (Pyrex); H, graded seal (Vitreosil-Pyrex); I, to vacuum line 11; J, Kovar; K, to mms spectrometer ionization chamber; L, to temperature controller; M, leak; N, cover; 0, reaction vessel; P, furnace; Q, Pt resistance element; MS, mass spectrometer.
mixing of reactants, mercury manometer, Toepler pump, gas-condensing device equipped with a Pyrex 1. torr sec-', at a pressure difference of leak (3 X 760 torr and a t room temperature) to allow slow evaporation of liquid, and three 1000-cc volumes to store purified gases. The reactor was connected to the mass spectrometer by way of 60 cm of copper tubing, i.d. 6 mm, provided with a stainless steel high-vacuum valve (bore 12 mm) to the mass spectrometer ionization chamber, and separated from the vacuum line by a glass stopcock. The inlet system and the glasscopper tubing between the reactor and the mass spectrometer were joined to a separate vacuum line. The copper tubing and the stainless steel valve were bakeable, respectively, a t 150 and 130". S a m p l i n g Device. The actual size of the quartz leak can be estimated from graphs of ion beam intensity against reactor pressure, as shown in Figure 2 . Ion beam intensity and gas pressure in the reactor show a linear relation over the whole pressure range. However, there is a change of slope at about 10 torr, very probably cor(6) E. A. Th. Verdurmen and C. A. Bank, J . Inorg. Nucl. Chem., 25, 1521 (1963). (7) Though it has been shown in the preliminary investigation6 t h a t small amounts of water do not influence the exchange rate, in the present experiments water has been carefully removed from the reactor, the inlet system, and the reactants.
Volume 71, Number 9 February 1967
E. A. TH.VERDURMEN
680
mean square velocity of the molecules, and f(7) is a complicated function of y = CJC, (the ratio of the specific heats of the gas a t constant pressure and a t constant volume). This kind of sampling by a mass spectrometer through a leak a t pressures for which the mean free path of the molecules is small compared to the diameter of the leak, has been considered theoretically and experimentally by Barber, Farren, and Linnett.9 According to their treatment, for a part,icular two-component gas mixture the ratio of the number of molecules per unit volume, nl and 5-22, in the mass spectrometer is given by
Peak height M/e 28
-[I
2-
- = pi Mi "' ni PZ Mz
n2
1-
0
7 I
0
where pl and p 2 are the partial pressures, and M1 and M z the molecular weights of the components in the reactor. If ion beam intensity I , concentration n, and ionization efficiency S, are related by I = Sn, then the following expression can be derived
CO pressure (torr) I
I
20
I
40
I
I
60
Figure 2. Peak height m/e 28 in the mass spectrum, as a function of CO pressure in the reactor, a t room temperature.
responding to the transition from molecular to viscous flow through the leak. Consequently, the size of the leak was of the order of the mean free path a t a pressure of 10 torr a t room temperature, Le., about 5 X mm. The total volume of the tubing on the low-pressure side of the leak was about 30 cc. Under operating conditions (gas pressure in the reactor 75 torr, and flow 3 x' 10-5 1. torr sec-l), the pressure just behind the leak could be estimated to be of the order of 10-3 torr. Thus, the total amount of gas in the tubing was equal t o the flow through the leak in about 0.5 sec. Pressure reduction in the reactor as a result of the continuous sampling through the leak was about 0.5% in 45 min. Consequently, the sampling did not disturb the kinetics. The response of the system showed some delay. Pressure changes in the reactor mass spectrometrically were detected after about 1 sec. Peak heights approached their ultimate values, within 1% of initial difference, after about 30 sec. Since concentration changes in the present experiments were slow relative to these 30 sec, they could be followed adequately in this system. The shape of the leak was that of a divergent The (maximum) mass rate of flow, above the criticalpressure ratio, is given by
G'
= ApCf(y)
(1)
where A is the area of the leak, p is the density of the gas on the high-pressure side of the leak, C is the root The Journal of Physical Chemistry
I n the present investigation, measured isotope ratios in CO and O2 were compared with a standard, so that fractionation effects cancelled out. However, product C 0 2was determined relative to the original amount of CO, and no standard was present, so that a mass discrimination correction had to be applied (vide infra). Materials. Natural carbon monoxide was prepared from purified carbon dioxide by reaction on charcoal a t 1000". Natural oxygen was made by gentle heating of KMnOa. A small amount of lsO2 was obtained by electrolysis of H2I80 (90%). Carbon dioxide, sulfur hexafluoride, nitrogen, helium, argon, and xenon were purest available materials. Xitric oxide was prepared by the action of moderately concentrated nitric acid on copper. Carbon monoxide and oxygen were purified by repeated condensation in a liquid nitrogen trap a t - 196", followed by slow evaporation collecting a middle fraction. Carbon dioxide, nitric oxide, and sulfur hexafluoride were purified by repeated bulbto-bulb distillation. Nitrogen was passed through a column of charcoal a t -196". The remaining gases were used without further treatment. The purity of the gases was checked by mass spectrometric analysis. In none of the samples was molecular hydrogen detected, indicating that the concentration of H, in the (8) A. B. Cambel and B. H. Jennings, "Gas Dynamics," McGrawHill Book Go., Inc., New York, N. Y., 1958, Chapter 7. (9) M. Barber, J. Farren, and J. W. Linnett, Proc. Roy. SOC.(London), A274, 293 (1963).
MECHANISM OF OXYGENISOTOPIC EXCHANGE IN CO-O2 MIXTURES
samp1.s was below 3 X The mass spectrum of the residual gas in the ionization chamber always showed a rather small peak of H20+. Thus the mass spectrometric analysis was not suitable for the detection of traces of HzO in the samples. The present drying method is similar to that of Gordon and Knipe.l0 These investigators observed that addition of 0.004% H 2 0 to their dried CO-O2 mixtures had a marked effect upon the position of the second explosion limit. Consequently, the concentration of H2O in their dried It mixtures must have been smaller than 1 X seems fairly safe to conclude that the present drying method provides samples containing less than 0.01% HzO. Other impurities (mainly nitrogen) did not exceed 0.1%. Treatment of Vessel. Formation of carbon dioxide in this system (below 600") appears to be mainly a surface process," At room temperature, the silica surface is believed to be covered with OH groups and possibly with physically adsorbed water molecules.12,13 Heating at 120" during some hours removes completely the physically adsorbed water, but the OH groups are retained. When the temperature is raised, water begins to be lost by formation of oxygen bridges between two adjacent hydroxyl groups. According to Zhadanov14 and de Boer, et aZ.,l3these oxygen bridges can be converted back to OH groups by exposure to liquid water. It has been observed that the catalytic formation of COz from CO and 0, at a quartz surface is the higher the less the number of hydroxyl groups attached to the silica surface. l 1 Therefore, to suppress the COz formation, before mounting, the reaction vessel was treated with water at 95" during 5 hr. Thus, total and partial pressure changes due to conversion of CO and O2 into C 0 2 and interference of product CO, were reduced as far as possible. At 500" generally not more than a few per cent of the maximum possible conversion into C02 was measured after 45 min of reaction time. The crushed silica (Vitreosil), used in a number of experiments, was treated in a similar way. Procedure. The water-treated reaction vessel, mounted into the furnace and connected to the vacuum line and the inlet system, was baked for 3 hr at 120" under vacuum. Then the temperature was raised to 500" and the vessel was evacuated for a t least 10 hr torr, measured by an ionization to below 1 X gauge. Before admitting a gas mixture, the vessel torr a t the was evacuated for 2 hr to below 1 X reaction temperature (between 500 and 580"). The reactants were premixed, except in the experiments with nitric oxide. This gas was admitted before or simultaneously with the remaining reactants, or it was added to a reacting mixture as indicated in the ex-
681
periments involved. Admission of reactants occurred by expansion from the premixing vessel. The gas pressure in the reactor was measured by a mercury manometer connected to the premixing vessel. The experiments were performed with natural CO and enriched 02. The oxygen sample used (l*Ocontent about 3%) was obtained by mixing natural oxygen and oxygen containing 90% l80isotope. The enrichment of the O2 thus predominantly was in the l80l8O species, so that next to isotopic exchange between CO and 0, the rate of attainment of statistical equilibrium of the l80in O2 (scrambling) could be measured. Recording of peak heights, corresponding to concentrations of isotopic CO, isotopic 0 2 , and isotopic COz molecules, generally was started about 40 sec after admission of reactants to the reactor. The range m/e 2 8 4 6 was continuously scanned during 45 min, so that the corresponding concentrations were measured direct)ly as a function of time. Before each series of experiments, pure oxygen gas was passed through the ionization chamber, by way of the conventional mass spectrometer inlet system, for 2 hr to condition the ion source. After that time, the ratios of the peak heights m/e 28 to m / e 32, and m/e 44 to m/e 32, appeared to be practically constant and relatively sniall (see Reduction of Data). A p p l i e d Rate Expressions. In this study it is assumed that the isotopic exchanges are ideal; thus isotope effects are not considered. The well-known first-order rate expression for isotopic exchange of the general type
AX,
+ BX,-lX*
J_ AX,-lX*
+ BX,
where X* and X are isotopes of a particular element, is given by
s = --1 t
nmab
na 4- mb
In (1 - F )
(3)
S is the gross rate of exchange; a and b are the partial pressures of the reactants AX, and BX,; F = (xt zo)/(z, - zo) is the exchange fraction; x t is the fractional content of tracer in AX, or BX, at time t, and zo and x, are the values for this fraction a t t = 0 and t = m , The kinetics of the present isotopic exchange reactions are complicated by a nonrandom distribution of isotopes in one of the reactants. However, it is (10) A. S. Gordon and R. H. Knipe, J . P h y s . Chem., 59, 1160 (1955). (11) C. A. Bank and E. A. Th. Verdurmen, ihid.,67, 2869 (1963). (12) J. C. Greaves and J. W. Linnett, Trans. Faraday Soc., 5 5 , 1355 (1959). (13) J. H. de Boer, M. E. A. Hermans, and J. M. Vleeskens, Proc. K o n i n k l . Ned. A k a d . Wetenschap., B60, 45 (1957); B61, 85 (1958). (14) S. P. Zhadanov, Dokl. Akad. S a u k S S S R , 6 8 , 99 (1949).
Volume 71, Number 3 February 1067
E. A. TH.VERDURMEN
682
easily derived that also in this case the general expression (3) can be applied. The general expression relating the rate of a reaction and the partial pressures of the reactants is
kaqbT (4) wherein q and r can be integers or fractional numbers. The following symbols are used in this study: (i) for formation of CO2, rate SCO,,pressure a = PO,, pressure b = p c o ; (ii) for oxygen isotopic exchange between CO and 02,rate SCO-O~ abbreviated SCO,exchange fraction Fco-o~, abbreviated FCO, pressure a = poZ, pressure b -= pco, n = 2, and m = 1; (iii) for oxygen exchange between 0 2 and 02 (scrambling), rate So2-o2 abbreviated S,,, exchange fraction Fo2-02 abbreviated F,,, pressure a = b = poZ,n = m = 2. Reduction of Data. The peak heights m/e 28 and 30 were corrected for the contribution of oxygen and carbon dioxide. These corrections were relatively small (a few per cent, rarely 10%). The addition of some of the foreign gases (N2, Ar, SF,) forced further correction at, respectively, in/e 28; m / e 36; and m / e 32,34, and 36. The starting point of the reaction in all performed experiments was arbitrarily fixed on the time t = 2 min after expanding the reactants into the reactor. From the corrected peak heights at m/e 28 and 30, 1 8 0 contents in CO were calculated, and hence the corresponding exchange fractions FCO were found as a function of time. From the corrected peak heights at m/e 32, 34, and 36, the product of mass ratios The 3*/32 and 34/36 was obtained as a function of time. oxygen samples used (l8O content about 3%) showed initial values of the product of about 0.005. The limiting value at equilibrium is 4.0. The product values were reduced to exchange fractions, F,,,using calculated curves of the product of mass ratios 34/32 and 34/36 as a function of exchange fraction. The measurement of the scrambling of isotopes in O2 was complicated by the decrease in '*O content of 02, due to the simultaneous isotopic exchange between 0 2 and CO. However, the latter exchange was relatively slow. The decrease in I 8 0 content of 0 2 generally was a few per cent, rarely lo%, of the initial value after 45 min reaction time. The cbxchange fractions F,, were corrected foy this decrease, assuming that the 3 4 0 2 and the 3 6 0 2 content in O2 are equally reduced as a result of the CO-O2 exchange. The amount of product COz was determined from the ratio of the corrected peak height m/e 44 to the corrected peak height m/e 28, the latter correlating with the CO partial pressure in the reactor. The ratio was corrected for the difference between the ionization efficiencies of C o and Coz, S
=
The Journal of Physical Chemistry
and a mass discrimination correction was applied (vide supra).
Results In this system three different processes occur: (a) formation of COZ; (b) isotopic exchange between CO and 0 2 ; (c) scrambling of oxygen isotopes in O2 (ie.,isotopic exchange between O2 and 0,). Figure 3, a typical plot of log (1 - F ) us. reaction time t, demonstrates that both the isotopic exchange CO-02 and the scrambling of oxygen isotopes in O2 fit in with expression 3: the relation between In (1 F ) and the reaction time t is obviously linear. For all performed experiments such linear relations have been found.15 Consequently, it is justified to calculate the exchange rates SCO and S,, from the slopes of the straight lines in the log (1 - F ) os. t plots, by substitution into expression 3. The rates SCO~, SCO,and S,, have been measured as a function of 0 2 pressure at constant CO pressure, and as a function of CO pressure a t constant O2 pressure a t 500". Plots of log S vs. log p can provide information on the magnitude of the numbers q and r in expression 4 and the applicability of the expression with certain q and r values over a particular pressure range. Such plots for the experiments mentioned are shown in Figures 4 and 5 . I n the latter, measurements have been included performed in a reaction vessel partly packed with crushed silica (Vitreosil) increasing the surface to volume ratio fivefold. The following observations are shown in these figures. (1) There is a striking difference in order of magnitude between (I-F)
Reaction time (min)
Q2
I
I
I
I
1
I
I
(15) Except for experiments in which very small amounts of NO were present (uide Figure io).
683
MECHANISM OF OXYGENISOTOPIC EXCHANGE IN CO-OZ MIXTURES
s ( torr
S(torr min-1)
4
1
4
rnin-'1
3
/
ip
10-'
SCO2
1o-2 ~
1
10-4
-
/
CO pressure ( t o r r )
1o 3
0, pressure ( t o r r )
1
100
10
Figure 5. Log-log plot of reaction rate 8 us. CO pressure a t constant pressure of 0 2 (24 torr) a t 500'; A and A, 0 and 0 , Sco; 0 and ., ScoZ; shaded marks: surface to volume ratio increased fivefold.
S C O ~SCO, , and S,,, a t each applied pressure. (2) It is obvious that none of the curves shown, except that of SCO,in Figure 5 , can be described by a particular expression (4)with constant numbers p and r over the whole pressure range. (3) Increasing the surface to volume ratio fivefold results in a roughly fivefold increase of SCO~, whereas SCOand s,, remain virtually constant. (4) The observed maxima in the curves Sco and S,, (Figure 5 ) are shifted to higher partial pressures of CO, if the surface to volume ratio is increased. Considering the nature of the occurring processes, observation 1 is a strong indication that these processes are independent. At each particular pressure in Figures 4 and 5 , the number of CO2 molecules formed per unit of time is much smaller than the number of exchanges CO-02 per unit of time, the latter number being much smaller than the number of exchanges 02-02 per unit of time. So the measured scrambling of oxygen isotopes in 0 2 cannot occur as a result of oxygen isotopic exchange between CO and Oz, nor can the latter exchange occur as a result of C 0 2 formation. Observation 3 confirms the conclusion of the preliminary study6 that the COZformation is mainly a surface
process, whereas the isotopic exchange reactions occur in the gas phase. From the shape of the curves, observation 2, it can be concluded that the isotopic exchange processes do not occur by the simple gasphase reactions ~ 1 6 0 160160
+
180160
+c 1 5 0
+
180180
--f
+
160180
160160
+
160180
(5) (6)
Some experiments have been performed studying scrambling of oxygen isotopes in pure oxygen. Even a t 600", practically no scrambling was measured within the reaction time (45 min). However, as is clearly shown in Figure 6, addition of a small amount (2%) of CO immediately starts scrambling a t a considerable rate a t 500". This observation indicates that the reactive species in the scrambling process is formed in a reaction in which CO is involved. The temperature dependence of the three processes has been determined by performing experiments a t various temperatures between 500 and 580". The measured rates of both isotopic exchange reactions and the measured rate of C 0 2 formation are logarithmically plotted as a function of reciprocal temperature, for two different mixtures, in Figure 7. The straight lines are a least-squares fit to experimental data. From Volume 71 Xumber 3 February 1967
E. A. TH.VERDURMEN
684
(1-F)
+ 1 t o y CO
Table I : Over-all Apparent Activation Energies for Scrambling in 0 2 , Isotopic Exchange CO-O,, and Formation of Con' Apparent activation energy, -kcal/mole------, Mixture Ib Mixture IIc
Process
0.4
Scrambling in 02 Isotopic exchange CO-02 Formation of COZ
j Reaction time (rnin)
0.2
4
I
I
I
I
28.5 1 2 . 3 34.011.4 30.0 & 2 . 5
29.5 f 1 . 4 36.4&1.8 36.7 f 5 . 5
a Observed in CO-02 mixtures in a quartz vessel between 500 and 580". Mixture of CO (24 torr) and 02 (24 torr). Mixture of CO (48 torr) and 02 (24 torr).
'
I
formation remains practically constant. It is easily found from the data reported in Table I1 that the ratio Ssc/Sc~does not depend upon foreign-gas pressure.
Table 11: Value of f(S) = S,,[CO]/~co[O~] as a Function of Foreign-Gas Pressure, for Addition of Various Foreign Gases"
i
Foreigngas pressure,
r
torr
He
Ar
0 12 24 48 72
8.04 7.87 7.40 7.35 7.57
..,b
8.25 7.96 8.21 8.16
f(S) for addition of Xe Na
7.36 7.56 7.29 7.15 ...b
7.89 7.99 7.89 7.71 769
a Observed in mixtures of CO (24 torr) and quartz vessel, at 500'. Not measured.
115
120
I
I
I
125
130
135
Figure 7. Logarithmic plot of reaction rate S vs. reciprocal temperature: A and A, SBc; 0 and 0 , Sco, 0 and ,. Scoz; open marks: CO (24 torr), 0 2 (24 torr); shaded marks: CO (48 torr), 0 2 (24 torr).
the slopes of these lines, apparentI6 activation energies have been obtained, that are reported in Table I. The apparent activation energies obtained do not show much difference, though the value relating to the isotopic exchange CO-O2 appears to be higher than the remaining apparent activation energies (about 30 kcal/ mole) by a few kcal/mole. I n order to determine the possible occurrence of excited intermediates with lifetimes of kinetic significance, experiments have been performed in which considerable amounts of foreign gases were added: He, Ar, Xe, Nz, COz, and SFs. The striking effect of these additions is that and co-02 isotopic exchange are retarded practically equally, whereas the rate of COa The Journal of Physical Chemistry
Cor ..,*
7.72 8.19 7.65 7.14 0 2
SFe
7.74 8.15 8.21 7.23 ...b
(24 torr) in a
Plots of the reciprocal of SCOus. foreign-gas pressure and plots of the reciprocal of S,, us. foreign-gas pressure show essentially straight lines. The rates Xcop, Scot and S,,, measured in a series of part'icular foreigngas additions, have been compared to the rates in the original CO-02 mixture that were measured preceding the- series. Though, generally, the reproducibility of SCOand S,, within a series of experiments was quite good (about f 5%), a comparison of the rates in similar experiments of different series showed larger scattering (about f 35%). Therefore, in order to compare the particular effects of the added gases, the intercepts of the straight lines obtained in the plots of reciprocal S us. foreign-gas pressure had to be reduced to one particular Of # at foreign-gas pressure* The particular effects Of the added gases are shown in Figures 8 and 9, being plots of reciprocal SCO and reciprocal S,, us. foreign-gas pressure. In (16) Apparent, since the particular activation energies obtained do not necessarily apply to an individual reaction.
685
MECHANISM OF OXYGEN ISOTOPIC EXCHANGE IN CO-O2 MIXTURES
1/sCO~02
si'
l/Solo8 ( A r b i t r a r y units)
(Arbitrary units)
/
Foreign gas pressure (torr)
0 0
20
I
I
I
40
60
80
Foreign gas p r e s s u r e ( t o r r )
Figure 8. Reciprocal of the rate of isotopic exchange CO-02 plotted as a function of foreign-gas pressure, for various foreign gases, CO (24 torr) and OZ (24 torr) at 500". Straight lines are least-squares fit to experimental data; slope x lo2: 0,He, 1.46 It 0.03; 0 , Ar, 1.16 & 0.34; V, Xe, 1.68 f 0.05; A, Nz, 3.88 f 0.16; 0 , COz, 3.97 f 0.45; W, SFs, 22.7 & 6.2.
Table I11 the rate of COe formation a t different foreigngas pressures, for addition of various foreign gases, has been given. Considering that within a series of experiments the reproducibility of SCO,appeared to be about f15%, it can be concluded from the data of Table I11 that the rate of C02 formation in these exTable I11 : Rate of ConFormation at Different Foreign-Gas Pressures, for Addition of Various Foreign Gases" Foreigngas pressure,
torr
--Sco2,
(torr min-1) X 108, for addition -fo Ar Xe N2
Re
'
0
3.1
...
7.6
12
3.8
8.0
24 48 72
4.2 4.2 4.2
4.9 5.3 6.0 6.4
9.1 9.6
. . .'
' Observed in mixtures of CO (24 torr) and quartz vessel, at 500'. Not measured.
'
4.9 6.0 6.0 6.0 6.0 0 2
SFa
6.9 3.1 2.7 2.2
. . .b
(24 torr) in a
0 , 0
I
20
LO
I
1
60
80
-
Figure 9. Reciprocal of the rate of scrambling in OZplotted as a function of foreign-gas pressure, for various foreign gases; CO (24 torr), O2 (24 torr) at 500'. Straight lines are least-squares fit to experimental data; slope X lo2: 0,He, 1.62 =t0.07; 0, Ar, 1.14 f 0.32; V, Xe, 1.85 i 0.03; A, Nz, 4.09 i 0.26; 0 , COz, 5.44 f 0.18; W, SFe, 36.6 f 20.8.
periments was practically constant. (The effect of SF6needs some qualification.) Figures 8 and 9 demonstrate the observed linear relation between the foreigngas pressure and the reciprocals of the exchange rate SCO and &. Moreover, it is obvious from these figures that the various foreign gases are not equally effective. Relative efficiencies have been calculated from the slopes of the straight lines compared to the slope of the straight line for N B addition. These efficiencies and their standard deviations are reported in Table IV. The efficiency of Nzhas been set equal to unity owing to the supposition that the effect of N2 does not differ appreciably from the foreign-gas effect of CO, a main component of the mixture. Taking into account the standard deviations mentioned, Table IV shows that the relative efficiencies for retardation of scrambling hardly differ from the relative efficiencies for retardation of isotopic exchange CO-02. Since Volume 71, Number 3 February 1967
E. A. TH. VERDURMEN
686
Isotopic exchange
Foreign gas
He Ar Xe
NZ
con
SFe
CO, P r e s s u r e
(1-F)
Table IV : Relative Efficiencies of Various Foreign Gases in the Retardation of the Scrambling in O2 and the Isotopic Exchange CO-Oz53b Scrambling in 0%
0.40 f 0.03 0.28 f 0.08 0.45 i 0 . 0 3 1.00 f 0.09 1 . 3 3 i 0.10 8.9 f 5 . 1
co-02 0.38 f 0.02 0.30 i.0.09 0 . 4 3 f 0.02 1.00 f 0.06 1.02 i 0.12 5 . 8 zk 1 . 6
- 1.5 - 1.0
a Observed in mixtures of CO (24 torr) and OZ(24 torr) in a quartz vessel at 500". * The efficiency of Nz has been set equal to unity.
the slope of the N2 line in Figure 8 practically equals the slope of the N2 line in Figure 9, respectively (3.88 f 0.16) X lo-? and (4.09 f 0.26) X the former observation further illustrates the conclusion that the scrambling in 0 2 and the isotopic exchange CO-02 are retarded practically equally by addition of foreign gases. If the retardation of exchange rates, as a result of added foreign gases, is produced by deactivation of excited intermediates, it is not probable that these gases mill have quite the same effect on different intermediates. Therefore, it is more likely that the main effect of addition of foreign gases is reducing the partial pressure of a common intermediate of both scrambling and CO-02 isotopic exchange. The almost equal efficiencies of He and Xe, and generally the order and relative magnitude of the efficiencies of the particular foreign gases, clearly suggest that neither adsorption nor diffusion in this system are rate-controlling processes. If oxygen atoms are an intermediate in this system, the addition of an 0 atom scavenger should inhibit the otherwise occurring processes in which 0 atoms are involved. In order to test this hypothesis, some experiments have been performed in which small amounts of NO were added. The third-order reaction of NO with 0 atoms is a fast and consequently efficient reaction. Therefore, KO is considered to be an efficient 0 atom scavenger. Generally, S O retards all chair reactions in which paramagnetic radicals propagate the reaction chains. Figures 10 and 11 are combined plots of log (1 - F ) us. reaction time and of pressure of product COz us. reaction time. In the experiments shown in Figure 10, the reactants, CO, 02, and NO, were admitted to the reactor simultaneously. In the experiments of Figure 11, NO was added to a reacting mixture roughly a t t = 21 min and t = 30 min. Both figures illustrate the following significant observations. T h e Journal of Physical Chemistry
2
10
20
30
LO
Figure 10. Logarithmic plot of (1 - F ) us. reaction time and plot of pressure of product COZus. reaction time for mixtures of CO (48 torr) and 0 2 (24 torr) and various amounts of NO a t 500'; 0, Fco; A, Fsc; 0,pco,. +02 tor,r NO
(1-F)
CO,
*3,torr NO
Pressure ( t o r r )
10
08
06
0.L
O2
p
1 I
I LO5
I
I
Reaction time ( m i n ) ,
2
10
20
30
I
,
LO
/
l
1 '
00
50
Figure 11. Logarithmic plot of (1 - F ) us. reaction time and plot of pressure of product COS us. reaction time for mixtures of CO (48 torr) and 0 2 (24 torr) at 500'. N O added a t times indicated. 0, Fco; A, F s c ; 0,peon.
(1) Small amounts of NO (concentrations roughly 3%) sate to inhibit both scrambling and CO-02 isotopic exchange completely. ( 2 ) Very small amounts of NO (concentration roughly 0.3%) markedly accelerate both scrambling and GO-02 isotopic exchange. (3) The rate of C o nproduction remains constant by the addition of very small amounts of KO, but at higher NO concentrations it is considerably enhanced. Regarding Figure 11, it has to be considered that the nitric oxide was introduced through the reactants inlet
MECHANISM OF OXYGENISOTOPIC EXCHANGE IN CO-02 MIXTURES
so that it took some time before mixing was complete. Hence, after addition of the second amount of NO, a t t = 30 min, the ultimate inhibition is preceded by acceleration of both isotopic exchange processes. Figure 10 clearly demonstrates that a t very small NO concentrations no linear relation is found between In (1 - 3') and the reaction time t , for both scrambling and CO-02 isotopic exchange. Obviously, there is no difference in the effect of YO whether it is added simultaneously with the remaining reactants or is introduced to a reacting mixture. Therefore, NO does not act upon the initiating reactions, but rather interferes in the propagating steps of the reaction process." Apart from its accelerating effect a t very small concentrations, the general behavior of KO in this system-complete inhibition of exchange processes by a relatively small amount of XO-is that of chain breaking by the inhibitor. As mill be argued in the Discussion section, the inhibitory effect of NO addition should be ascribed to very small amounts of KO2, produced by slow reaction of the added S O and the 0 2 reactant.
centrations. On the contrary, the observation that a relatively small amount of NO completely inhibits the exchange processes is a strong argument for occurrence of a chain mechanism involving free atoms in very small concentrations. Since, additionally, oxygen atoms are generally believed to occur as important intermediates in this system a t slightly higher temperaSOO"), it seems justified to assume that t u r e ~ ' - (above ~ oxygen atoms act as the common intermediate in the present processes too. The scrambling of oxygen isotopes in O2and the COOr isotopic exchange observed in this system thus are ascribed to the independent processes
0 0
+
0 2
+0 2
+0
(7)
+ co -+co + 0
(8)
Termination and Initiation. Termination of chains involving oxygen atoms can occur by wall recornbination of 0 atoms, reaction 9, or by the gas-phase thirdorder reactions 10 and 11
+ 0 + w + + 'ct' 0 + oz + 11 --+ + 11: 0 + co + 31 +COZ + 31 0
Discussion Reactive Species. The main information about the nature of the reactive species is obtained from the effect of the added inert foreign gases and the effect of S O addition. It has been concluded that the main foreigngas effect is reducing the partial pressure of a common intermediate of both scrambling and CO-02 isotopic exchange, Considering the nature of both isotopic exchange processes, it is likely that this intermediate will be one of the species 02,0 3 , 0 2 * , and 0.'* The shape of the curves in Figures 4 and 5 eliminates O2 as the common intermediate: reactions 5 and 6 have been rejected. Os shows rapid reaction with NO,19 but it has the disadvantage that its concentration is not expected to be reduced by the addition of inert gases. Recently, some chemical reactions of the metastable lowest-lying known excited states of oxygen, Or( 'Ag) and Oz(l2,+), have been studied.20-22 KOrapid reaction with NO was ~ b s e r v e d . ~ l - ~Apart * from their long radiative lifetime, these states appear to be very stable with respect to molecular collisions. Winer and BayesZ5observed at 23" and at a pressure of 10 torr, in a 20-mm i.cl. reaction tube, that wall deactivation of O2(lAP) completely predominates gas-phase deactivation, It has been found in the present investigation that the concentration of the intermediate is considerably reduced by the addition of foreign gas at pressures below 100 torr. Vibrationally excited 0 2 in the electronic ground state a t 500" probably will be easily deactivated by collision.26 Thus it cannot set up long chains and, therefore, should occur in rather high con-
687
0 2
(9)
0 3
(10) (11)
The direct recombination in the gas phase is too slow to be of importance. Indicating the rate of 0 atom formation by Ro, a steady-state treatment yields the following expression for the 0 atom concentration in the stationary state LO1
=
Ro
k-lO[OZl[l\lI
+ k-11[COI[Ml+
k9
(12)
(17) V. N. Kondratev, "Chemical Kinetics of Gas Reactions," Pergamon Press Ltd., Oxford and London, 1964, p 606. (18) The intermediates Cot*, CzO, and CdOt, probably occurring in this system at slightly higher temperatures,32~are all produced by reactions involving 0 atoms. (19) H. S. Johnson and H. J. Crosby, 6.Chem. Phys., 22, 689 (1954). (20) R. E. March, S. G. Furnival, and H. I. Schiff, Photochem. Photobiol., 4, 971 (1965). (21) M . A. A. Clyne, B. A. Thrush, and R. P.Wayne, ibid., 4, 957 (1965). (22) L. 11'. Bader and E. A. Ogryzlo, Discussions Faraday Soc., 37, 46 (1964). (23) S. J. Arnold, E. A. Ogryzlo, and H. Witzke, J . Chem. Phys., 40, 1769 (1964). (24) L. Elias, E. A. Ogryzlo, and H. I. Schiff, Can. J . Chem., 37, 1680 (1959). (25) A. M .Winer and K. D. Bayes, J . Phys. Chem., 70,302 (1966). (26) Lipscomb, et al., measured a t room temperature very different efficiencies for deactivation of vibrationally excited 0%( 2 = 8) in the electronic ground state: COz, one in 7000 collisions; Ar and Nz, one in more than lo7 collisions. In the present study, the efficiencies of the various foreign gases are relatively in the same order of magnitude (the efficiencies of Nz and CO, are almost equal). F. J. Lipscomb, R. G. 1%Norrish, '. and B. A. Thrush, Proc. Roy. Soe. (London), A233, 456 (1956).
Volume 71, Siimber 3
Febiuaru 1967
E. A. TH.VERDURMEN
688
wherein wall recombination of oxygen atoms has been regarded as a first-order process, according to Greaves and Linnett.12 The rate constant ke can be calculated from the recombination coefficient of oxygen atoms on Vitreosil,12 y (500") = 1.4 X lo+, and the vessel dimensions: ks (at 5 0 0 O ) = 355 sec-'. The rate constant klo[M = 0 2 1 can be obtained from data by Benson and A ~ w o r t h y , ?klo ~ = 2.96 X lo7 exp(890/ R T ) 1.2 mole-2 sec-l, a t 500" reduced to klo = 1.5 X cm6 molecule-2 sec-l. The rate constant kll can be calculated from data by Kondratev and Ptichkin,28kll (at 428°K) = 2.96 X 1013cma mole-2 sec-l, and data by Clyne and regarding the activation energy of t,he reaction: 3.7 kcal/mole; kll (at 500") = 5.6 X cm6 molecule-2 sec-l. Thus, at 500" klo and kll are in the same order of magnitude. I n order to determine the order of magnitude of the 0 atom concentration in the present experiments, it is assumed that the measured rate of scrambling is given by the rate of the isotopic exchange reaction 0
+
0 2
+0 2
+0
(74
The rate constant of this reaction has been determined by Herron and Klein30and by Brennen and Niki.31 Assuming the reaction has the same small negative activation energy as reaction 10, the rate constant a t 500" is given by = 0.4 X 10-l2 cm3molecule-l sec-l The rate of reaction 7a is represented by S 7 a = k7&[021[01
(13)
In a mixture of CO(24 torr) and 0 2 (24 torr) a t 500", a scrambling rate S,, = 0.379 torr min-I, i.e., 8.8 X 1013 molecules C M - ~ sec-l has been measured. By substitution into expression 13, the concentration of 0 atoms is found to be [O] = 6.3 X lo8 atoms cm-3. Then the rate of 0 atom formation, Ro, required to maintain this concentration is calculated from expres- ~ sion 12: Ro = 3.2 X IO1' atoms ~ m sec-'. As is shown in Figure 6, scrambling of oxygen isotopes in O2 occurs under condition that CO is present, so that obviously the reactive intermediates, 0 atoms, are formed in a reaction in which CO is involved. This reaction might be
c0+0?+C02+0
(14)
Sulzmann, et al.,32determined the rate constant of this reaction in the temperature range 240C-3000°K, in a shock-wave experiment k14
=
(3.5 f 1.6) X lo1?exp[-((51000 f 7000)/RT] cm3 mole-1 sec-1
Hence, it is expected that the over-all apparent activaThe Journal of Physical Chemistry
tion energies in the present system will exceed 44 kcal/ mole. The actual apparent activation energies reported in Table I evidently are much smaller than the 51 f 7 kcal/mole observed by Sulzman in reaction 14. Consequently, the gas-phase reaction 14 can be rejected as an important source of 0 atoms in this system. The most likely alternative is a similar process (15) occurring a t the wall of the reaction vessel
CO+Oz=CO2+O (15) It is shown in Figure 5 that the rate of COzformation is proportional to the surface to volume ratio of the vessel. Consequently, the COz production is mainly a surface process, and it is not unlikely that the process (15) is involved. I n the mixture of CO (24 torr) and O2 (24 torr) a t 500", a C02 formation rate SCO,= 0.012 torr min-l, ie., 2.8 X 10l2 molecules cm-3 sec-l, has been measured. If the COS formation is attended with the simultaneous production of an 0 atom, the experimentally determined value of Scot should equal the rate of 0 atom formation. However, it can be expected that only part of the produced 0 atoms actually will escape from the wall into the gas phase. I n the mixture of CO (24 torr) and 0 2 (24 torr), the deduced value of ROis about 11% of the measured rate SCO,. If chains are started at the wall of the vessel, the observation that the isotopic exchange rates S,, and SCoremain virtually constant in spite of the enhancement of the surface to volume ratio demonstrates that chain breaking a t the wall should be important too. Considering expression 12, both Ro and ks are proportional to the surface to volume ratio. By substitution of the rate constants ks, klo, and kll, it is found that, in a mixture of CO (24 torr) and O2(24 torr), a fivefold enhancement of the surface to volume ratio will increase the 0 atom concentration by a factor of 1.3. Actually, this effect had to be measured by comparison of experiments in a reaction vessel partly packed with crushed silica, with similar experiments in an unpacked vessel. Since the reproducibility of S,, and SCOin such experiments of different series appeared to be about =t35%, the expected small enhancement of the 0 (27) S. W. Benson and A. E. Axworthy, J . Chem. Phys., 42, 2614 (1965). (28) V. N. Kondratev and I. I. Ptichkin, Kinetika i Kataliz, 2, 492 (1961). (29) M . A. A. Clyne and B. A. Thrush, Proc. Roy. SOC. (London), A269, 404 (1962). (30) J. T. Herron and F. S. Klein, J . Chem. Phys., 40, 2731 (1964); 41, 1285 (1964); 44, 3645 (1966). (31) W. Brennen and H. Niki, ibid., 42, 3725 (1965). (32) K. G. P. Sulsmann, B. F. Myers, and E. R. Bartle, ibid., 42, 3969 (1965).
MECHANISM OF OXYGEN ISOTOPIC EXCHANGE IN CO-02 MIXTURES
atom concentration could not be measured. The observed experimental scattering should have its origin in the activity of the surface that obviously is not invariable. The condition of the wall, e.g., the number of OH groups at the silica surface, will determine not only the rate of C 0 2production," and consequently the rate of 0 atom formation, but also the fraction of 0 atoms that leaves the wall. Moreover, the less OH groups a t the silica surface the higher will be the recombination coefficient for oxygen atoms on the wa11.12 Thus the dependence of the 0 atom concentration in the gas phase upon the surface conditions is rather complicated. summarizing, the initiation and wall termination of chains in this system is interpreted in terms of the reactions
co + Owat1
c02
0 2
+
Owall
---+
+
Owsll
(16)
(02)wa~
(17)
(18)
O w a l l +Ogas
+ +surface- + surface- + O,,, +surfaceO*
surface-O*
Ogas
(02),,,
(19)
rapid
(20)
The mechanism of wall recombination of oxygen atoms on Vitreosil, reactions 19 and 20, has been discussed by Greaves and Linnett.12 Oxygen atoms from the gas phase recombine with loosely bound oxygen atoms, 0*, that have been adsorbed from the gas phase. Propagation. Isotopic exchange of oxygen atoms with 0 2 has been studied by Herron and Klein.30 Their results have been interpreted in terms of the mechanism
0
+
0 2
os*
oe*+ M -+oa+ n1
(21, 21')
(22)
The rate constants kZ1' and kZ2were found to be a t room temperature, kzl' = 1.8 X lo9 sec-l and k22 = 4 X 10" cm3 mole-l sec-'; i.e., 0.7 X 10-l2 cma molecule-' sec-l. The stationary-state concentration of 0 3 is given by
and the rate of exchange by
szlf =kzl~[~a*~
(24)
Assuming that no temperature corrections in k21' and kn have to be applied, it is easily calculated that a t 500" and total pressures below 100 torr, holds:
689
kz2[Ml/k21'
